Calibration and registration of camera arrays using a single circular grid optical target
Methods for determining intrinsic parameters associated with a camera and for registering cameras within a camera array are described. In some embodiments, a camera may be calibrated using an optical target that is positioned at a depth outside the depth of focus of the camera. The optical target may include a grid of symmetrical shapes (e.g., a rectilinear grid of circles or tilted squares). The intrinsic parameters for the camera may be determined using an iterative calibration process in which a cost function is used to evaluate the straightness of lines intersecting rows and columns associated with centroids of the symmetrical shapes. In some embodiments, the registration of a color camera with a depth camera may include mapping centroids identified within a first color image captured by the color camera with corresponding centroids identified within an undistorted intensity image captured by the depth camera.
1. A method for calibrating a camera array, comprising:
capturing a first image of an environment including an optical target using a first camera, the optical target includes a plurality of symmetrical shapes;
determining a plurality of centroid locations within the first image corresponding with the plurality of symmetrical shapes;
determining one or more distortion parameters associated with the first camera based on the plurality of centroid locations;
generating a reduced distortion image using the first image and the one or more distortion parameters;
outputting the reduced distortion image;
capturing a second image of the optical target using a second camera different from the first camera;
determining a registration mapping for the second image onto the reduced distortion image;
generating a registered image based on the second image and the registration mapping; and
outputting the registered image.
2. The method of claim 1 , wherein:
the optical target is positioned within the environment at a depth outside a depth of focus of the first camera.
3. The method of claim 1 , wherein:
the plurality of symmetrical shapes includes a plurality of circles.
4. The method of claim 1 , wherein:
the determining a plurality of centroid locations includes thresholding the first image and determining a center of mass for each of the plurality of symmetrical shapes.
5. The method of claim 1 , further comprising:
determining a center of distortion based on the one or more distortion parameters; and
determining a principal point based on the center of distortion.
6. The method of claim 1 , further comprising:
determining an average distance between neighboring centroids within the reduced distortion image; and
determining a focal length associated with the first camera based on the average distance between neighboring centroids.
7. The method of claim 1 , wherein:
the determining one or more distortion parameters associated with the first camera includes optimizing the one or more distortion parameters using a cost function that evaluates the straightness of lines intersecting rows and columns associated with the plurality of centroids.
8. The method of claim 1 , further comprising:
determining depth information associated with the reduced distortion image, the first camera comprises a depth camera, the registration mapping includes one or more registration mapping coefficients; and
adjusting the one or more registration mapping coefficients based on the depth information.
9. The method of claim 1 , wherein:
the outputting the reduced distortion image includes displaying the reduced distortion image;
the first camera comprises a depth camera; and
the second camera comprises a color camera.
10. A system for calibrating a camera array, comprising:
a depth camera, the depth camera to capture a first image of an environment including an optical target, the optical target includes a plurality of symmetrical shapes;
one or more processors, the one or more processors to communicate with the depth camera, the one or more processors to determine a plurality of centroid locations within the first image corresponding with the plurality of symmetrical shapes, the one or more processors to determine one or more distortion parameters associated with the depth camera based on the plurality of centroid locations, the one or more processors to generate a reduced distortion image using the first image and the one or more distortion parameters; and
a color camera to communicate with the one or more processors, the color camera to capture a second image of the environment including the optical target, the one or more processors to determine a registration mapping for the second image onto the reduced distortion image, the one or more processors to generate a registered image by applying the registration mapping to the second image.
11. The system of claim 10 , wherein:
the optical target is positioned within the environment at a depth outside a depth of focus of the depth camera.
12. The system of claim 10 , wherein:
the plurality of symmetrical shapes includes a plurality of circles.
13. One or more storage devices containing processor readable code for programming one or more processors to perform a method for calibrating a camera array comprising the steps of:
capturing a first image of an environment including an optical target using a depth camera, the optical target includes a plurality of symmetrical shapes;
determining a plurality of centroid locations within the first image corresponding with the plurality of symmetrical shapes;
determining one or more intrinsic parameters associated with the depth camera based on the plurality of centroid locations;
generating a reduced distortion image using the first image and the one or more intrinsic parameters;
capturing a second image of the optical target using a color camera;
determining a registration mapping for the second image onto the reduced distortion image;
generating a registered image using the second image and the registration mapping; and
outputting the reduced distortion image and the registered image.
14. The one or more storage devices of claim 13 , wherein:
the optical target is positioned within the environment at a depth outside a depth of focus of the depth camera.
15. The one or more storage devices of claim 13 , wherein:
the plurality of symmetrical shapes includes a plurality of circles.
16. The one or more storage devices of claim 13 , wherein:
the determining a plurality of centroid locations includes thresholding the first image and determining a center of mass for each of the plurality of symmetrical shapes.
17. The one or more storage devices of claim 13 , wherein:
the outputting the undistorted image and the registered image includes displaying the reduced distortion image.
18. The one or more storage devices of claim 13 , further comprising:
determining an average distance between neighboring centroids within the reduced distortion image; and
determining a focal length associated with the depth camera based on the average distance between neighboring centroids.